Azole derivatives of a general formula I, having the 2R,3R-configuration, wherein R1, R2 and R3 represent, independently from one another, hydrogen, fluorine, chlorine or CF3-group are valuable antifungal drugs for treatment of systemic mycoses and possess a excellent broad antifungal activity.

For example Isavuconazole [(2R,3R)-3-[4-(4-cyanophenyl)thiazol-2-yl)]-1-(1H-1,2,4-triazol-1-yl)-2-(2,5-difluorophenyl)-butan-2-ol; formula I, R1 and R3 represent fluorine and R2 represents hydrogen] as well as Ravuconazole [(2R,3R)-3-[4-(4-cyanophenyl)thiazol-2-yl)]-1-(1H-1,2,4-triazol-1-yl)-2-(2,4-difluorophenyl)-butan-2-ol; formula I, R1 and R2 represent fluorine and R3 represents hydrogen] are useful antifungal drugs as reported in U.S. Pat. No. 5,648,372 from Feb. 1, 1995 or in U.S. Pat. No. 5,792,781 from Sep. 18, 1996 or in U.S. Pat. No. 6,300,353 from Oct. 9, 2001 (WO99/45008).
Since compounds of general formula I contain two adjacent chiral centers, synthesis of enantiomerically pure compound is complex and until now, all patented syntheses are not efficient enough and do not allow cost effective manufacturing on a technical scale:
Thus, U.S. Pat. Nos. 5,648,372 or 5,792,781 describe enantioselective synthesis of compounds of formula I (specifically Ravuconazole) from chiral 3-hydroxy-2-methyl propionic acid in 12 steps with overall yield lower than 5%. In another approach including 13 steps and low overall yield, (R)-lactic acid was used as the starting material (Chem. Pharm. Bull. 46(4), 623 (1998) and ibid. 46(7), 1125 (1998)). Because both starting materials contain only one chiral center, in a number of inefficient steps, the second, adjacent chiral center has to be created by a diastereoselective reaction (using either Corey or Sharpless epoxidation method) which is not sufficiently selective leading mostly to a mixture of two diastereomers which have to be separated. The second approach, based on (R)-methyl lactate, was recently very thoroughly optimized by BMS on a multi kilogram scale but it still does not fulfill requirements for cost effective manufacturing process (Organic Process Research & Development 13, 716 (2009)). The overall yield of this optimized 11 steps process is still only 16% (Scheme 1).

The manufacturing process for Isavuconazole is similar: Since Isavuconazole differentiates from Ravuconazole by only another fluorine substitution on the aromatic ring (2,5- instead of 2,4-difluorophenyl), the identical synthesis has been used (U.S. Pat. No. 6,300,353 from Oct. 9, 2001 and Bioorg. & Med. Chem. Lett. 13, 191 (2003)). Consequently, also this manufacturing process, based on (R)-lactic acid, faces the same problems: to many steps, extremely low overall yield and in addition to U.S. Pat. No. 6,300,353 claims even already known step as novel (claim 36).
Recent attempts to improve this concept as reported in WO 2007/062542 (Dec. 1, 2005), using less expensive, natural configured (S)-lactic acid, also failed: As already reported in U.S. Pat. No. 6,133,485 and in US 2003/0236419, the second chiral center was formed from an optically active allyl alcohol prepared in a few steps from (S)-lactic acid. This allyl alcohol was subjected to Sharpless diastereoselective epoxidation providing first an opposite configured, epimeric epoxy alcohol which had to be then epimerized in an additional inversion step yielding finally the desired epoxy alcohol as the known precursor for Isavuconazole (U.S. Pat. No. 6,300,353). It is obvious that this process using less expensive (S)-lactic acid makes the entire process with an inversion step even more complex than the original approach.
Elegant and more efficient process has been claimed in US 2004/0176432 from Jun. 26, 2001) in which both chiral centers have been formed simultaneously, diastereo- and enantio-selectively pure in one single reaction step using chiral (R)-2-butynol as a chiral precursor in the presence of Pd(II)-catalyst and diethyl zinc (Scheme 2).

Since water soluble, (R)-2-butynol is expensive, recently identical process has been published, in which instead of (R)-2-butynol less water soluble and therefore, less expensive (R)-4-phenyl-3-butyn-2-ol was used (Synthetic Commun. 39, 1611 (2009)). Nevertheless, as incorrectly stated there, this process does not provide better diastereoselectivity than the original process using (R)-2-butynol: On the contrary disadvantage of this process is a very bad atom economy because huge phenyl group of (R)-4-phenyl-3-butyn-2-ol has to be “disposed” in oxidation step by the conversion of triple bond into carboxylic acid function.
All known processes for enantiomerically pure compounds of formula I have definitely too many operation steps and specifically very low overall yield. The chiral starting materials used, either 3-hydroxy-2-methyl propionic acid or (S)- or (R)-methyl lactate, contain only one chiral center and consequently, in number of steps, the second adjacent chiral center has to be ineffectively generated which makes the entire process long and expensive. The only known process, which generates both chiral centers simultaneously, requires again expensive chiral starting material (R)-2-butynol.